TPS54620 design issue

I don't see anything obvious in your schematic.  Can you verify that you have 5 V at the input )pins 4, 5 and 6)?  What is the VSENSE voltage?  Can you show some waveforms, including the PH node?

Dear John,

2654.HDCAST TEST2.pdf

I checked 5V input (PVIN1/PVIN2/VIN) is ok. As shown in the photo shots, there are normal and abnormal (not working) cases.When it is normal, the waveform is a square one as shown in the attachment.

When we power reset, it sometimes ends up with abnormal case where it does not generate 3.3V at all (2V or so). I wonder if C413 (680uF) may cause trouble when it is power-reset. I would appreciate your comments.

Regards,

Jai

Have you measured the PVIN/VIN pins? It's hard to tell for certain with the 2ms time scale, but the abnormal PH waveform looks like PVIN is dropping out after ~3ms of switching, and causing a UVLO.

Where is the input power coming from in your captures, a lab supply or upstream regulator. Are the R686 and R682 resistors placed? if so, what is the resistance values?

I would also suggest looking at Comp, Boot, and the inductor current for additional clues. 

Hi Matt,

The answers to your questions are:

1. PVIN/VIN = 5 V.

2. Input power 5V comes directly from HP power supply.

3. R486 and R682 = 0 ohm but not connected.

I wonder if you have any reference design. I would appreciate it.

Regards,

Jai